Overall, results indicated that there were high relationships between caregiver and self-report responses on measures of symptom severity, daily living skills, and quality of life. However, discrepancies between reporters on level of reported symptom severity impacted clinical utility despite the lack of statistically significant differences. Additionally, the combination of self-report and caregiver-report on all measures better predicted employment outcomes than did an individual reporter. In particular, caregiver report of daily living skills was valuable in determining the likelihood that an adult was currently employed. Taken together, these findings suggest that, while self-report is valid for this subset of adults with ASD, a multi- informant approach should be best clinical practice for assessment in this population.
Specifically, the present study investigated whether self-report and caregiver measures of symptom severity, adaptive daily living skills, and quality of life were: (1) significantly
discrepant from one another, (2) significantly positively correlated even if discrepancies existed, and (3) differentially predictive of employment and independent living outcomes. Exploratory analyses examined whether higher caregiver-reported symptom severity in the area of social communication was associated with more disagreement between caregiver-report and self-report on measures of daily living skills and quality of life.
A primary aim of this study was to elucidate the extent to which self-report on ASD symptoms can be reliably used for adults with ASD of average to above average intellectual functioning. In assessing symptom severity, our findings indicate that adults with ASD and their
also demonstrated positive significant correlation with one another. The correlation of r = .50 found in the present study was not significantly different from the average cross-informant correlation of r = .66 reported in the original standardization study of the SRS-A (Z = 1.46, p = .07; Constantino & Gruber, 2007). Further, neither self-report nor caregiver report of symptom severity proved more useful in predicting objective measures such as employment or need for daily supports. Although ASD symptom severity scores were, on average, remarkably consistent across reporters, discrepant reports resulted in different clinical classifications for 47.5% of the sample. Further, for 22.5% of pairs, caregiver scores placed the adult with ASD within the clinical range while self-report scores resulted in a sub-threshold classification.
Taken together, the present study indicates that, overall, adults with ASD of average to above average intellectual functioning can serve as reliable and accurate reporters of their own symptoms. These findings contrast with previous research suggesting that poor social insight limits the validity of self-report for this particular population (e.g., Berthoz & Hill, 2005; Bishop & Seltzer, 2012). Exploratory analyses did suggest that higher levels of caregiver-reported difficulties with social communication were associated with adults with ASD reporting more daily living skills and higher quality of life than their caregivers reported about them. While this finding may suggest that higher impairments in social insight in this population may lead to greater difficulties in accurately reporting daily living skills and quality of life, it may also driven by the fact that caregivers who reported higher levels of symptom severity were more likely to report more elevated difficulties in other areas, as well. Knowing that self-report in this
population is likely to provide valuable information about symptom severity will allow self- report measures to become incorporated into comprehensive assessment methods. Recognizing the utility of self-report in this capacity can help to shape treatment planning, as well as provide a
means of measuring intervention effectiveness in research contexts. However, the fact that clinical classifications were not always consistent across reporters suggest that including a caregiver is still important in clinical contexts and a multi-informant approach is likely to
provide the most comprehensive information. These findings also suggest that it may be useful to establish different clinical cut-off scores for self-report measures than for informant-report measures in order to ensure that adults with significant levels of impairment are not being incorrectly disqualified from services.
Accurately assessing daily living skills is also a key part of shaping treatment planning and supporting increased independence for adults with ASD. In the present study, we found that, despite a significant positive correlation between scores, there was a significant discrepancy between caregiver and self-report scores on the W-ADL. Specifically, caregivers reported that the adults with ASD demonstrated fewer daily living skills on average than adults reported about themselves. This discrepancy may be driven in part by the framing of the W-ADL items, which ask about what skills the adult can do as well as what skills the adult does. This difference between what someone “can do” vs. what someone “actually does” may contribute to divergence in reporting between adults with ASD and caregivers. An outside reporter may be better able to objectively monitor what skills are actually conducted on a regular basis. Interestingly, however, the correlation between caregiver and self-report scores on the W-ADL was significantly higher than the average correlation observed in other populations, supporting research indicating that cross-informant agreement is typically higher for more visible or objective constructs
(Achenbach et al., 2005; G. Meyer et al., 2001)
Notably, caregiver-report on the W-ADL was the only measure found to significantly increase predictive accuracy of current employment status of the adult with ASD. This finding is
in line with longitudinal data showing that better daily living skills was one of the keys to successfully finding and maintaining employment for adults with ASD (Klinger et al., 2015). Because the addition of caregiver-reported daily living skills resulted in increased prediction of an objective outcome measure compared to self-report scores alone, caregiver-report may be more accurate in comprehensively assessing independence in completing daily living skills. Having accurate information on the W-ADL is essential in both clinical and research contexts, as adaptive behaviors and daily living skills have been consistently shown to be one of the best measures of long-term outcomes. For example, a study of adult outcomes for people with ASD and cognitive functioning in the average range found that adaptive behavior measures –
particularly in the daily living skills domain – were most closely correlated with outcomes in independent living, working, and social functioning (Farley et al., 2009). These findings also support research indicating that IQ scores alone are not always reliable prognostic indicators (Howlin et al., 2004, 2014; Klinger et al., 2015), as individuals with high IQ scores but low daily living skills had poorer independent living outcomes than individuals with relatively low IQs who were able to care for themselves with little assistance. Because our data showed that caregiver-report on the W-ADL provided additional information that was not gathered in the self-report assessment alone, daily living skills may be the area in which it is most essential to involve informant report.
With regards to assessment of quality of life, overall, caregivers reported significantly lower quality of life – particularly in the Satisfaction domain – on the QoL-Q than adults with ASD reported about themselves, despite a significant positive correlation between reporters. The finding that self-reported quality of life was higher than caregiver ratings is consistent with research in other populations (Ikeda et al., 2014; Shipman et al., 2011), though correlation
between reporters in our sample was actually higher than would be expected based on such research. Improving quality of life is a primary consideration in interventions and services for adults with ASD (Gerber et al., 2011); as such, an individual’s subjective feelings about one’s own life must be taken into account and, in many cases, may be given higher priority than an informant’s report. However, it is also important to consider the objective factors that influence quality of life, such as independence and success in work or social contexts. Klinger and
colleagues (2015) found that higher daily living skills predicted employment and that
employment was significantly associated with higher quality of in regards to satisfaction, sense of belonging, and empowerment. Because neither self-report nor caregiver report of quality of life proved more useful in predicting employment status, results of the current study do not provide evidence that one reporter is more accurate than another.
While a multi-informant approach has been a standard recommendation in the assessment of psychopathology in children, as well as adults with developmental or personality disorders (Barkley et al., 2011), this approach has not been well-studied for individuals with ASD.
Inclusion of a caregiver or other informant in the assessment of daily living skills is particularly essential, as our data showed that caregivers added significant information that was not provided by adults with ASD themselves. Because we know that daily living skills influence employment outcomes, which in turn have a significant impact on overall quality of life, improving daily living skills should be a primary target of interventions with adults (Duncan & Bishop, 2013; Hume et al., 2009; Klinger et al., 2015). As such, it is imperative to have an accurate picture of what an adult’s daily living skills truly look like.
Despite the demonstrated value of involving multiple informants in assessing adults with ASD, this study also shows that self-report should be considered an important consideration for
individuals with average or above intellectual ability. As self-report has not traditionally been part of assessments for ASD, there has been consequently very little research addressing the appropriateness of self-report measures in this population. This lack of research had impacted not only the provision of clinical services, but also the progress of research on adults with ASD. Particularly for the field of intervention research, it is impossible to move forward without knowing if self-report assessments can adequately capture the unique symptomatology of ASD in adulthood or measure treatment effectiveness. The present study demonstrates that, while a multi-informant approach is ideal, adults with ASD are capable of reporting on many aspects of their own lives. The fact that there was not a significant discrepancy between reporters on the SRS-A and that self-report on this measure was significantly predictive of employment outcomes indicates that adults with ASD are able to accurately report on their levels of symptom severity and ASD-related impairment. While there were significant discrepancies between reporters on both the W-ADL and QoL-Q, correlation between reporters was still on par with (or better than) interrater agreement in other populations. The combination of self-report on the SRS-A, W- ADL, and QoL-Q together was also predictive of employment outcomes, indicating that self- report accurately maps onto objective constructs. By identifying areas in which caregivers and adults with ASD are most likely to disagree, and the areas in which caregiver-report adds significant information that may not have been gathered from self-report alone, such as daily living skills, we can better guide decisions about if and when multiple informants should be consulted in making diagnoses and treatment plans. We can also use this information the help shape future assessment protocols and practices so that the field of intervention research for adults with ASD can continue to progress.
Limitations and Future Directions
Although this study makes significant strides in improving our understanding of self- report adults with ASD, there are still several limitations. Firstly, while this study was
sufficiently powered to conduct the present analyses, the sample size was still relatively small. Future research examining a larger sample is recommended. Additionally, the full sample was recruited from a pool of adults who were diagnosed as children at the TEACCH Autism
Program. Individuals who received diagnoses as children are often different from those who did not receive diagnoses until later in life, as more substantial symptoms often result in earlier diagnoses. Consequently, our findings may not be representative of self-report capabilities across the full autism spectrum. However, finding that adult self-report of ASD symptoms is consistent with caregiver report doesn’t support these sample concerns.
Future directions for this study include expanding participant range of age and
intellectual functioning. A large proportion of adult assessments for ASD are conducted around the transition age (i.e., late teens to early twenties); the average age in our sample (33.17 years) was older than that time period, and the results presented here may be less applicable to
transition-aged adolescents and young adults. Additionally, because our sample was recruited from childhood records, IQ cutoffs for inclusion were also based on reported childhood IQ; thus, it is difficult to characterize the current cognitive functioning of adults in our sample. Given that there may be larger discrepancies in scores for individuals with lower IQs, future studies would benefit from having access to current IQ scores and should aim to include individuals from a wider range of intellectual functioning.
Finally, analyses involving the W-ADL should also be interpreted with caution, as the measure contains only 17 items and the range of scores is 1 to 34. Given the limited range of
possible scores, correlation between reporters was likely inflated as both members of adult- caregiver pairs hit ceiling in many instances. Future research including a standardized measure of adaptive behavior/independent living skills is needed to confirm the findings in this study.
Follow-up analyses of the present sample could also include examining the extent to which reporters differ specifically on what adults can do vs. what they actually do.
The overarching goal of this research was to provide a greater understanding of what information different reporters provide on existing measures, as well as how particular measures map onto outcomes in order to help to shape assessment practices and treatment planning.
Moving forward, this research aims to help shape future measures as they are being developed so that clinicians and researchers are able to ascertain the most accurate picture of how an
individual is functioning in the areas of symptom severity, daily living skills, and quality of life. For example, this may mean having different cut-off scores for self-report than for informant report for areas in which we know the reporters differ (such a daily living skills), and next steps include investigating if and how such cut-off scores should be established. In a research context, having access to valid self-report measures is also essential to move forward with
treatment/intervention studies in this population, as such measures are essential reliably
capturing changes from pre- to post-intervention. The present study may serve as a first step in demonstrating the validity of self-report in this population, allowing intervention research to move forward using self-report as a meaningful assessment tool.
Summary
Overall, this study promotes a multi-informant approach as best clinical practice for assessing various aspect of ASD, including symptom severity, daily living skills, and quality of life. While results indicated remarkable consistency across reporters for assessing autism
symptomatology, when differences occurred they resulted in different clinical classifications for 47.5% of the sample. Thus, a multi-informant approach is recommended for assessments
conducted to either diagnose ASD or to evaluate symptom severity. Further, a multi-informant approach is supported by findings that both caregiver- and self-report of daily living skills improved the prediction of employment outcomes. Thus, when evaluations are being conducted to recommend adult independent living and employment support services, results suggest that it is important to include both self- and caregiver-reports. Overall, when it comes to qualification for services and shaping treatment plans, the inclusion of both caregiver- and self-report may ensure a more comprehensive picture of current functioning across domains.
Table 1. Demographics for the total sample (n = 40).
Sex (%male) 80% (n = 32)
Mean Age (SD; range) 33.17 (5.54; 23.83 – 47.84)
% Caucasian 87.5% (n = 35) Employment Status (% employed) 57.5% (n = 23) Living Situation With Family 57.5% (n = 23) Independently 30% (n = 12) Supervised Housing 12.5% (n = 5) Caregiver (% mothers) 72.5% (n = 29) *p ≤ .05
Table 2. Differences between self-report and caregiver-report on included measures. Adult: Mean (SD) Caregiver: Mean (SD) Test Statistic (df) Significance (two-tailed) SRS-A Total (t-scores) 60.26 (9.49) 61.97 (12.25) t(37) = -.95 p = .35
SCI 59.92 (9.01) 60.97 (12.27) t (37)= -1.17 p = .25 RRB 63.20 (12.06) 65.03 (12.78) t(39) = -.80 p = .43 QoL-Q Total (3 domains) 67.57 (7.67) 65.08 (8.87) F(2, 72) = 3.40 p = .04*
Satisfaction 22.43 (3.49) 20.55 (4.08) t(39) = 2.96 p = .002* Belongingness 19.89 (3.78) 19.95 (4.37) t(36) = 1.14 p = .93 Empowerment 25.13 (3.60) 24.78 (3.79) t(39) = 1.28 p = .45 W-ADL Total 30.00 (3.81) 28.87 (4.39) t(37) = 2.36 p = .02*
Figure 1.Correlation between caregiver and self-report total scores (t-scores) on the SRS-A. Higher scores indicate greater symptom severity. A t-score of 59 and below is classified as “within normal limits,” 60 to 65 as “mild range” 66 to 75 as “moderate range,” and 76 or higher as “severe range.” Reference line (red) represents r = 1.00.
Figure 2.Correlation between caregiver and self-report total scores on the W-ADL. Scores range from 0 to 34. Higher scores indicate a greater number of daily living skills used independently. Reference line (red) represents r = 1.00.
Figure 3. Correlation between caregiver and self-report scores on the QoL-Q across three subdomains: Satisfaction, Belongingness, and Empowerment. Higher scores indicate higher quality of life. Reference line (red) represents r = 1.00.